Identifying Key Words in 9-1-1 Calls for Stroke: A Mixed Methods Approach

Emergency Medical Services dispatcher recognition of stroke: A systematic review

PURPOSE

Stroke treatments are time-sensitive, and thus early and correct recognition of stroke by Emergency Medical Services is essential for outcomes. This is particularly important with the adaption of mobile stroke units. In this systematic review, we therefore aimed to provide a comprehensive overview of Emergency Medical Services dispatcher recognition of stroke.

METHODS

The review was registered on PROSPERO and the PRISMA guidelines were applied. We searched PubMed, Embase, and Cochrane Review Library. Screening and data extraction were performed by two observers. Risk of bias was assessed using the QUADAS-2 instrument.

FINDINGS

Of 1200 papers screened, 24 fulfilled the inclusion criteria. Data on sensitivity was reported in 22 papers and varied from 17.9% to 83.0%. Positive predictive values were reported in 12 papers and ranged from 24.0% to 87.7%. Seven papers reported specificity, which ranged from 20.0% to 99.1%. Six papers reported negative predictive value, ranging from 28.0% to 99.4%. In general, the risk of bias was low.

DISCUSSION

Stroke recognition by dispatchers varied greatly, but overall many patients with stroke are not recognised, despite the initiatives taken to improve stroke literacy. The available data are of high quality, however Asian, African, and South American populations are underrepresented.

CONCLUSION

While the data are heterogenous, this review can serve as a reference for future research in emergency medical dispatcher stroke recognition and initiatives to improve prehospital stroke recognition.

The Burden of Stroke Mimics Among Hyperacute Stroke Unit Attendees with Special Emphasis on Migraine: A 10-Year Evaluation

BACKGROUND AND OBJECTIVES

Stroke and migraine are common neurological illnesses that cause tremendous suffering for patients. Certain diseases can mimic the clinical manifestations of an actual stroke. Migraine is one of the most commonly reported stroke mimics. The main goals of this study are to look at the prevalence of stroke mimics on the stroke pathway of Sheffield Teaching Hospitals and how many of them are migraines.

MATERIALS AND METHODS

A retrospective service evaluation was conducted at the hyperacute stroke unit (HASU) of the Royal Hallamshire Hospital (RHH) in the United Kingdom. The total admissions from 2013 to 2022 were collected from the Sentinel Stroke National Audit Programme database, and the number of stroke mimics was evaluated each year. The burden of migraine stroke mimics was also evaluated. Then, a one-year sample of stroke mimics was extracted to look for the types of each mimic.

RESULTS

From 2013 to 2022, 45.75% (n = 12156) of the stroke pathway patients (n = 26573) were stroke mimics, with an increment of up to 55% in the years 2021 and 2022. During these 10 years, migraine stroke mimics accounted for 10.21% of admissions (n = 1240). The three most common mimics in a one-year sample of stroke pathway patients were migraine (14.70%) (n = 373), functional neurological disorders (FNDs) (7.17%) (n = 182), and Guillain-Barré syndrome (6.66%) (n = 169). Seizures, syncope, and metabolic derangements were reported as mimics in 4.17% (n = 106), 3.14% (n = 80), and 1.77% (n = 45), respectively.

CONCLUSIONS

About half of the HASU attendees were stroke mimics rather than actual strokes, and the most common mimics were migraines.

Telephone triage and dispatch of ambulances to patients with suspected and verified acute stroke - a descriptive study

OBJECTIVES

In this study we aimed to explore EMCC triage of suspected and confirmed stroke patients to gain more knowledge about the initial phase of the acute stroke response chain. Accurate dispatch at the Emergency Medical Communication Center (EMCC) is crucial for optimal resource utilization in the prehospital service, and early identification of acute stroke is known to improve patient outcome.

MATERIALS AND METHODS

We conducted a descriptive retrospective study based on data from the Emergency Department and EMCC records at a comprehensive stroke center in Oslo, Norway, during a six-month period (2019-2020). Patients dispatched with EMCC stroke criteria and/or discharged with a stroke diagnosis were included. We identified EMCC true positive, false positive and false negative stroke patients and estimated EMCC stroke sensitivity and positive predictive value (PPV). Furthermore, we analyzed prehospital time intervals and identified patient destinations to gain knowledge on ambulance services assessments.

RESULTS

We included 1298 patients. EMCC stroke sensitivity was 77% (95% CI: 72 - 82%), and PPV was 16% (95% CI: 14 - 18%). EMCC false negative stroke patients experienced an increased median prehospital delay of 11 min (p < 0.001). Upon arrival at the scene, 68% of the EMCC false negative patients were identified as suspected stroke cases by the ambulance services. Similarly, 68% of the false positive stroke patients were either referred to a GP, out-of-hours GP acute clinic, local hospitals or left at the scene by the ambulance services, indicating that no obvious stroke symptoms were identified by ambulance personnel upon arrival at the scene.

CONCLUSIONS

This study reveals a high EMCC stroke sensitivity and an extensive number of false positive stroke dispatches. By comparing the assessments made by both the EMCC and the ambulance service, we have identified specific patient groups that should be the focus for future research efforts aimed at improving the sensitivity and specificity of stroke recognition in the EMCC.

Prehospital Stroke Care Part 1: Emergency Medical Services and the Stroke Systems of Care

Acute stroke care begins before hospital arrival, and several prehospital factors are critical in influencing overall patient care and poststroke outcomes. This topical review provides an overview of the state of the science on prehospital components of stroke systems of care and how emergency medical services systems may interact in the system to support acute stroke care. Topics include layperson recognition of stroke, prehospital transport strategies, networked stroke care, systems for data integration and real-time feedback, and inequities that exist within and among systems.

A mixed methods analysis of caller-emergency medical dispatcher communication during 9-1-1 calls for out-of-hospital cardiac arrest

OBJECTIVE

Recognition of out-of-hospital cardiac arrest (OHCA) during 9-1-1 calls is critically important, but little is known about how laypersons and emergency medical dispatchers (EMDs) communicate. We sought to describe 9-1-1 calls for OHCA.

METHODS

We performed a mixed-methods, retrospective analysis of 9-1-1 calls for OHCA victims in a large urban emergency medical services (EMS) system using a random sampling of cases containing the term "cardiopulmonary resuscitation" (CPR) in the EMS electronic report. A constant comparison qualitative approach with four independent reviewers continued until thematic saturation was achieved. Quantitative analysis employed computational linguistics. Callers' emotional states were rated using the emotional content and cooperation score (ECCS).

RESULTS

Thematic saturation was achieved after 46 calls. Three "OHCA recognition" themes emerged [ 1) disparate OHCA terms used, 2) OHCA mimics create challenges, 3) EMD questions influence recognition]. Three "CPR facilitation" themes emerged [ 1) directive language may facilitate CPR, 2) specific instructions assist CPR, 3) caller's emotions affect CPR initiation]. Callers were generally "anxious but cooperative." Callers saying "pulse" was associated with OHCA recognition.

CONCLUSION

Communication characteristics appear to influence OHCA recognition and CPR facilitation.

PRACTICE IMPLICATIONS

Dispatch protocols that acknowledge characteristics of callers' communication may improve OHCA recognition and CPR facilitation.

Machine Learning-Based Text Analysis to Predict Severely Injured Patients in Emergency Medical Dispatch: Model Development and Validation

Background

Early recognition of severely injured patients in prehospital settings is of paramount importance for timely treatment and transportation of patients to further treatment facilities. The dispatching accuracy has seldom been addressed in previous studies.

Objective

In this study, we aimed to build a machine learning–based model through text mining of emergency calls for the automated identification of severely injured patients after a road accident.

Methods

Audio recordings of road accidents in Taipei City, Taiwan, in 2018 were obtained and randomly sampled. Data on call transfers or non-Mandarin speeches were excluded. To predict cases of severe trauma identified on-site by emergency medical technicians, all included cases were evaluated by both humans (6 dispatchers) and a machine learning model, that is, a prehospital-activated major trauma (PAMT) model. The PAMT model was developed using term frequency–inverse document frequency, rule-based classification, and a Bernoulli naïve Bayes classifier. Repeated random subsampling cross-validation was applied to evaluate the robustness of the model. The prediction performance of dispatchers and the PAMT model, in severe cases, was compared. Performance was indicated by sensitivity, specificity, positive predictive value, negative predictive value, and accuracy.

Results

Although the mean sensitivity and negative predictive value obtained by the PAMT model were higher than those of dispatchers, they obtained higher mean specificity, positive predictive value, and accuracy. The mean accuracy of the PAMT model, from certainty level 0 (lowest certainty) to level 6 (highest certainty), was higher except for levels 5 and 6. The overall performances of the dispatchers and the PAMT model were similar; however, the PAMT model had higher accuracy in cases where the dispatchers were less certain of their judgments.

Conclusions

A machine learning–based model, called the PAMT model, was developed to predict severe road accident trauma. The results of our study suggest that the accuracy of the PAMT model is not superior to that of the participating dispatchers; however, it may assist dispatchers when they lack confidence while making a judgment.

Stroke mimics: incidence, aetiology, clinical features and treatment

Mimics account for almost half of hospital admissions for suspected stroke. Stroke mimics may present as a functional (conversion) disorder or may be part of the symptomatology of a neurological or medical disorder. While many underlying conditions can be recognized rapidly by careful assessment, a significant proportion of patients unfortunately still receive thrombolysis and admission to a high-intensity stroke unit with inherent risks and unnecessary costs. Accurate diagnosis is important as recurrent presentations may be common in many disorders. A non-contrast CT is not sufficient to make a diagnosis of acute stroke as the test may be normal very early following an acute stroke. Multi-modal CT or magnetic resonance imaging (MRI) may be helpful to confirm an acute ischaemic stroke and are necessary if stroke mimics are suspected. Treatment in neurological and medical mimics results in prompt resolution of the symptoms. Treatment of functional disorders can be challenging and is often incomplete and requires early psychiatric intervention.

Help-seeking behaviour and subsequent patient and system delays in stroke

OBJECTIVES

Prehospital delay is the main reason why only a limited number of stroke patients receive reperfusion therapy. We aimed to investigate help-seeking behaviour in patients and bystanders after onset of stroke and subsequent patient and system delay.

MATERIALS & METHODS

We conducted a cross-sectional study of 332 patients with stroke. We performed structured interviews and used data from the medical records and the Danish Stroke Registry. Primary outcomes were patient delay and system delay.

RESULTS

The median patient delay was 280 min, and the median system delay was 97 min. For a patient delay of <3 h, an additional non-significant system delay of median 30 min was seen for a first contact to a general practitioner (GP), and an additional significant delay of median 490 min was seen for the small group of patients with a first contact to 'other' healthcare professionals compared to the Emergency Medical Services (EMS). For a patient delay of more than 3 h, an additional system delay of median 78 min was found when the first contact was directed to the out-of-hours primary care (OOH-PC). A total of 17% of patients were admitted to another hospital or department before arrival at the stroke centre; this resulted in a substantially prolonged system delay of a median of 431 min.

CONCLUSIONS

Patient delay remains the main reason for delayed arrival at the stroke centre. Appropriate help-seeking behaviour and efficient pre-hospital triage are essential for reducing the prehospital delay and increasing the proportion of patients receiving reperfusion therapy.

Early recognition of a caller's emotion in out-of-hospital cardiac arrest dispatching: An artificial intelligence approach

AIM

This study aimed to develop an AI model for detecting a caller's emotional state during out-of-hospital cardiac arrest calls by processing audio recordings of dispatch communications.

METHODS

Audio recordings of 337 out-of-hospital cardiac arrest calls from March-April 2011 were retrieved. The callers' emotional state was classified based on the emotional content and cooperative scores. Mel-frequency cepstral coefficients extracted essential information from the voice signals. A support vector machine was utilised for the automatic judgement, and repeated random sub-sampling cross validation (RRS-CV) was applied to evaluate robustness. The results from the artificial intelligence classifier were compared with the consensus of expert reviewers.

RESULTS

The audio recordings were classified into five emotional content and cooperative score levels. The proposed model had an average positive predictive value of 72.97%, a negative predictive value of 93.47%, sensitivity of 38.76%, and specificity of 98.29%. If only the first 10 seconds of the recordings were considered, it had an average positive predictive value of 84.62%, a negative predictive value of 93.57%, sensitivity of 52.38%, and specificity of 98.64%. The artificial intelligence model's performance maintained preferable results for emotionally stable cases.

CONCLUSION

Artificial intelligence models can possibly facilitate the judgement of callers' emotional states during dispatch conversations. This model has the potential to be utilised in practice, by pre-screening emotionally stable callers, thus allowing dispatchers to focus on cases that are judged to be emotionally unstable. Further research and validation are required to improve the model's performance and make it suitable for the general population.

Symptoms presented during emergency telephone calls for patients with spontaneous subarachnoid haemorrhage

Background

A spontaneous subarachnoid haemorrhage (SAH) is one of the most critical neurological emergencies a dispatcher can face in an emergency telephone call. No study has yet investigated which symptoms are presented in emergency telephone calls for these patients. We aimed to identify symptoms indicative of SAH and to determine the sensitivity of these and their association (odds ratio, OR) with SAH.

Methods

This was a nested case–control study based on all telephone calls to the medical dispatch center of Copenhagen Emergency Medical Services in a 4-year time period. Patients with SAH were identified in the Danish National Patient Register; diagnoses were verified by medical record review and their emergency telephone call audio files were extracted. Audio files were replayed, and symptoms extracted in a standardized manner. Audio files of a control group were replayed and assessed as well.

Results

We included 224 SAH patients and 609 controls. Cardiac arrest and persisting unconsciousness were reported in 5.8% and 14.7% of SAH patients, respectively. The highest sensitivity was found for headache (58.9%), nausea/vomiting (46.9%) and neck pain (32.6%). Among conscious SAH patients these symptoms were found to have the strongest association with SAH (OR 27.0, 8.41 and 34.0, respectively). Inability to stand up, speech difficulty, or sweating were reported in 24.6%, 24.2%, and 22.8%. The most frequent combination of symptoms was headache and nausea/vomiting, which was reported in 41.6% of SAH patients. More than 90% of headaches were severe, but headache was not reported in 29.7% of conscious SAH patients. In these, syncope was described by 49.1% and nausea/vomiting by 37.7%.

Conclusion

Headache, nausea/vomiting, and neck pain had the highest sensitivity and strongest association with SAH in emergency telephone calls. Unspecific symptoms such as inability to stand up, speech difficulty or sweating were reported in 1 out of 5 calls. Interestingly, 1 in 3 conscious SAH patients did not report headache.

Trial registration NCT03980613 (www.clinicaltrials.gov).

Enhanced dispatch and rendezvous doubles the catchment area and number of patients treated on a mobile stroke unit

INTRODUCTION

Mobile Stroke Units (MSUs) deliver acute stroke treatment on-scene in coordination with Emergency Medical Services (EMS). One criticism of the MSU approach is the limited range of a single MSU. The Houston MSU is evaluating MSU implementation, and we developed a rendezvous approach as an innovative solution to expand the range and number of patients treated.

METHODS

In addition to direct 911 dispatch of our MSU to the scene within our 7-mile catchment area, we empowered more distant EMS units to activate the MSU. We also monitored EMS radio communications to identify possible patients. For these distant patients, the MSU met the EMS unit en route to the stroke center and treated the patient at that intermediate location. The distribution of the distance from MSU base station to site of stroke and time from 911 alert to tissue plasminogen activator (tPA) bolus were compared between patients treated on-scene and by rendezvous using Wilcoxon rank sum test.

RESULTS

Over 4 years, 338 acute ischemic stroke patients were treated with tPA on our MSU. Of these, 169 (50%) were treated on-scene after MSU dispatch at a median of 6.4 miles (IQR 6.4 miles) from MSU base station. 169 (50%) were treated by 'rendezvous' pathway with assessment and treatment of stroke a median of 12.4 miles from base (IQR 5.5 miles) (p< 0.0001). Time (min) from MSU alert to tPA bolus did not differ: 36.0 ± 10.0 for on-scene vs 37.0 ± 10.0 with rendezvous (p=0.65). 13% of patients alerted via direct 911 dispatch were treated vs 44% of rendezvous patients.

CONCLUSION

Adding a rendezvous approach to an MSU dispatch pathway doubles the range of operations and the number of patients treated by an MSU in an urban area, without incurring delay.

Utstein recommendation for emergency stroke care

Background

Recent advances in treatment for stroke give new possibilities for optimizing outcomes. To deliver these prehospital care needs to become more efficient.

Aim

To develop a framework to support improved delivery of prehospital care. The recommendations are aimed at clinicians involved in prehospital and emergency health systems who will often not be stroke specialists but need clear guidance as to how to develop and deliver safe and effective care for acute stroke patients.

Methods

Building on the successful implementation program from the Global Resuscitation Alliance and the Resuscitation Academy, the Utstein methodology was used to define a generic chain of survival for Emergency Stroke Care by assembling international expertise in Stroke and Emergency Medical Services (EMS). Ten programs were identified for Acute Stroke Care to improve survival and outcomes, with recommendations for implementation of best practice.

Conclusions

Efficient prehospital systems for acute stroke will be improved through public awareness, optimized prehospital triage and timely diagnostics, and quick and equitable access to acute treatments. Documentation, use of metrics and transparency will help to build a culture of excellence and accountability.

Régulation médicale : mise au point 2018 et développements futurs

En 2018, les centrales de régulation d’urgences médicales devraient toutes être en mesure d’offrir certaines plus-values essentielles aux patients : une réanimation par téléphone avec accompagnement des témoins jusqu’à l’arrivée des secours, l’utilisation de réseaux de premiers répondants, ainsi que les défibrillateurs automatiques des domaines privés et publics via des applications Internet, l’assistance aux témoins pour réaliser des injections d’épinéphrine lors de réactions allergiques sévères ou de chlorhydrate de glucagon lors d’hypoglycémie. En complément, le bénéfice attendu des centrales d’urgences médicales visà-vis du système de soins dans son ensemble implique une meilleure efficience quant à l’engagement des moyens dont elles disposent. Cela ne peut se faire qu’en collectant les données liées à l’engagement des moyens (type, nombre, niveau de priorité) et en les corrélant avec la gravité de la situation observée sur site par les intervenants. Cette analyse permet d’estimer les taux de sur-triage et de sous-triage de la centrale lors de la décision d’envoi des moyens. L’utilisation appropriée des signaux prioritaires qui en découle permet également de réduire les risques des intervenants et de la population générale. Finalement, les fusions de centrales permettent de rationaliser certains coûts fixes, de faire face à des difficultés de recrutement et de formation, mais également d’augmenter la qualité de la régulation offerte à la population.

Emergency Dispatcher Stroke Recognition: Associations with Downstream Care

OBJECTIVE

As the first point of contact for patients activating emergency medical services (EMS), emergency dispatchers have the earliest opportunity to recognize stroke. We sought to quantify dispatcher stroke recognition and its relationships with EMS stroke recognition and response speed.

METHODS

We assembled a cohort of consecutive EMS-transported patients with a dispatcher suspected stroke or a hospital discharge diagnosis of stroke or transient ischemic attack (TIA). Dispatcher sensitivity and positive predictive value (PPV) for stroke recognition were calculated. Multivariable logistic regression analysis was used to determine predictors of dispatcher recognition and relationships between dispatcher recognition and downstream care.

RESULTS

During a 12-month period, 601 patients met inclusion criteria. Dispatchers suspected stroke in 229/324 (sensitivity = 70.7% [65.5 to 75.4%]) confirmed stroke/TIA cases and correctly assigned a suspected stroke label in 229/506 cases (PPV = 45.3% [41.0 to 49.6%]). Dispatchers had higher odds of recognizing ischemic strokes (aOR 3.4 [1.4 to 8.5]) and lower odds of recognizing patients with visual deficits (aOR = 0.4 [0.2 to 0.9]) or vomiting (aOR = 0.3 [0.1 to 0.9]). Dispatcher suspected stroke cases received more on-scene stroke screens (79.0% vs. 54.7%, p < 0.0001) and were more often recognized by EMS as strokes (77.7% vs. 57.9%, p = 0.0005). Dispatcher recognition was independently associated with EMS stroke recognition (aOR = 3.8 [1.9 to 7.7]), but not with transportation times, door-to-CT times, or t-PA delivery.

CONCLUSIONS

Emergency dispatcher stroke recognition is associated with higher rates of on-scene stroke scale performance and EMS ischemic stroke recognition but not with reduced transport times, door-to-CT times, or t-PA treatment.